Forecasting dangerous driving to prevent crashes

Forecasting dangerous driving to prevent crashes
Share on
Key Takeaways

What if you knew when, where, and why collisions occur… before they happened?

Transportation planners commonly work with injury and fatality data to help set priorities for where and how to improve road safety. Although this is valuable information, it’s usually too late by the time they understand and act on it: people have already been hurt or even killed. This reactive approach is a paradox for achieving Vision Zero because it predicates the elimination of traffic deaths and injuries on collecting data about traffic deaths and injuries.

Zendrive is committed to helping experts get ahead of the curve and develop “upstream” approaches to prevent casualties. We do this by measuring and analyzing the driving behaviors that are most likely to contribute to crashes: speeding, aggressive driving and distracted phone use. If you can deter these dangerous behaviors, you can prevent injuries and fatalities.

Zendrive is using our 15-billion miles of data for predictive analytics that can tell us where collisions are likely to occur before they happen. In partnership with New York University’s Tandon School of Engineering, our massive amount of data is being put to work on a scale like never before. Both organizations are inspired by Mayor Bill de Blasio’s Vision Zero initiative, which aims to eliminate traffic deaths and serious injuries in New York by 2024.

NYU analyzed and mapped 33,450 risky driving events collected by Zendrive between July and December 2015 and 127,423 collisions reported by the NYPD between July 2012 and March 2017. During those four-plus years, over 1,200 people were killed in traffic on New York City streets. Zendrive and NYU entered into this partnership to find ways to accelerate the pace of saving lives on the roads.

What did we learn?

By analyzing and mapping the data, NYU researchers found mappable correlations between driver behavior data and NYPD crash data. Basically, they determined  that the areas where people drive recklessly are the same areas where there are collisions. This means that it is possible to stop reckless driving before it causes collisions, injuries and deaths.

For their analysis, NYU researchers matched specific driver behaviors measured by Zendrive to specific NYPD collision types:

  • Fast Acceleration
  • Aggressive Driving/Road Rage
  • Hard Braking
  • Following Too Closely
  • Phone Use While Driving
  • Cell Phone (hands-free),
  • Cell Phone (hand-held),
  • Driver Inattention/Distraction
  • Speeding above speed limit
  • Unsafe Speed

Then, they mapped the datasets and compared the locations and density of the events in each category.

Overall, researchers found a 71-percent overlap between the locations of dangerous driving behaviors and traffic crashes around New York City.

                 All dangerous driver behavior (Zendrive)                                                      All collisions (NYPD)

When it came to driver phone use and cell-phone-related crashes, they found a 75-percent overlap.

                     Driver phone use (Zendrive)                                                      Phone-related collisions (NYPD)

The relationships between fast acceleration and aggressive driving/road rage collisions, along with  hard braking and following too closely collisions, were just below 70-percent each.

These are all strong findings that can be used to accelerate progress towards achieving Vision Zero. Policy-makers can act on them now, and researchers can continue to to learn more about saving lives faster.

Check out the New York Daily News’ take on the study.

What’s next?

There is lifesaving potential in comparing massive amounts of aggregated, anonymous driver behavior data to collision data (and other other datasets) on a large scale and in real-time.

The larger the dataset, the more precise and actionable steps communities can take to achieve Vision Zero. Today, Zendrive collects tens of millions of dangerous driving events a month in New York City, compared to the hundreds of injuries and less than 20 fatalities the NYPD records each month. While there are still far too many casualties, from an analytical-perspective, it’s a relatively small number compared to the millions of times people drive dangerously on a regular basis — all of which make people feel unsafe.

Real-time driving behavior data allows decision-makers and communities to assess traffic engineering and regulatory changes before, during and after project implementation. Real-time data can also improve education and enforcement campaigns. It can help  focus resources on deterring the behaviors that are most likely to result in collisions at the highest risk locations and on t the most crash-prone days and times. This data is invaluable to all Vision Zero stakeholders.

Fortunately, with this level of insight, city leaders no longer have to wait for crashes and casualties to occur on their roads. They can use the data to figure out when and where to deploy interventions that will deter specific behaviors that would otherwise lead to collisions, injuries and deaths.

Whether you call this approach “proactive,” “predictive,” “upstream” or “forecasting,” it’s all about mitigating the most dangerous behaviors before they reach their typical outcomes. With this strategy, we no longer have to rely on traffic fatality and injury data to prevent fatalities and injuries. Real-time “big data” provides faster, more accurate information that we can act on right now.

Contact our public policy and governmental affairs team if you want learn about how to work with Zendrive.

Get notified of new articles
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.